A non-cardiomyocyte autonomous mechanism of cardioprotection involving the SLO1 BK channel

Reagents, animals

NS11021 was a gift from Neurosearch A/S (Ballerup, Denmark). All other chemicals were of the highest grade available from Sigma (St. Louis, MO) unless otherwise specified. Slo1^−/− (KCNMA1^−/−) mice were generated on the FVB background via deletion of Exon1, which contains the translation start site and the S0 trans-membrane segment (Meredith et al., 2004). This results in both transcripts and protein being undetectable, and it has been previously shown that the lack of an S0 domain results in a nonfunctional BK channel  (Zarei et al., 2001; Zarei et al., 2004). Animals were genotyped by tail biopsy PCR, as previously described (Meredith et al., 2004). All experiments used male WT and Slo1^−/− littermates age 6–8 weeks. Mice were maintained in an AAALAC-accredited pathogen-free barrier facility with food and water available ad libitum. All procedures were in accordance with the NIH Guide for the Care and Use of Laboratory Animals (2011 revision) and were approved by the University Committee on Animal Resources.

Mouse Langendorff ex-vivo perfused heart

Following avertin anesthesia, a thoracotomy was performed, the aorta cannulated in-situ and the heart rapidly transferred to perfusion apparatus as previously described (Wojtovich et al., 2011; Nadtochiy et al., 2009). The heart was perfused with Krebs-Henseleit buffer in constant flow mode (4 ml/min). The following experimental protocols were observed: (i) IR injury: 20 min normoxic perfusion, 40 min global ischemia, 60 min reperfusion. (ii) NS1619+IR: 10 min normoxic perfusion, 10 min of 5 µM NS1619, 30 s washout, then IR as above. (iii) NS11021+IR: 10 min normoxic perfusion, 10 min of 500 nM NS11021, 30 s washout, then IR as above.

Studies involving hexamethonium were independently controlled and the following experimental protocols were observed: (iv) IR injury: 25 min normoxic perfusion, 35 min global ischemia, 60 min reperfusion. (v) NS1619+IR: 15 min normoxic perfusion, 10 min of 5 µM NS1619, 30 s washout, then IR as above. (vi) NS1619+IR+Hexamethonium: 12.5 min normoxic perfusion, 2.5 min of 500 µM hexamethonium, 10 min of 5 µM NS1619 plus 500 µM hexamethonium, 30 s washout of NS1619 only, 35 min global ischemia, 5 min reperfusion plus 500 µM hexamethonium, 55 min reperfusion. (vii) Atpenin A5 (AA5)+IR: 5 min normoxic perfusion, 20 min of 10 nM AA5, 30 s washout, then IR as above. (viii) AA5+IR+Hexamethonium: 2.5 min normoxic perfusion, 2.5 min of 500 µM hexamethonium, 20 min of 10 nM AA5 plus 500 µM hexamethonium, 30 s washout of AA5 only, 35 min global ischemia, 5 min reperfusion plus 500 µM hexamethonium, 55 min reperfusion. Note that different ischemic times (35 vs. 40 min) were used to ensure adequate independent controls for every group examined.

To assess neuronal survival, hearts were exposed to 100 µM nicotine for 5 min following the reperfusion period. Compounds were delivered via a syringe pump linked to an injection port immediately above the perfusion cannula. The volume of added solutions was ≤0.05% of the total perfusate volume. Following experimental protocols, hearts were TTC stained and imaged as previously described to quantify infarct size (Wojtovich et al., 2011; Nadtochiy et al., 2011).

Mouse cardiomyocyte isolation

Following avertin anesthesia, the aorta was cannulated in-situ and the heart rapidly transferred to perfusion apparatus. The heart was perfused for 3 min with a modified Krebs-Henseleit “buffer A” (118 mM NaCl, 4.7 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 10 mM HEPES, 24 mM NaHCO₃, 5.5 mM glucose, bubbled with 95% O₂, 5% CO₂, pH 7.35, 37 °C) at 3 ml/min to remove blood. The heart was then perfused with digestion buffer (buffer A plus 0.25 mM CaCl₂, 30 mM taurine, 10 mM 2,3-butanedione monoxime, 0.0278% Trypsin, 25 mg collagenase A, 75 mg collagenase D) at 3 ml/min for 12 min. The heart was removed and teased apart with forceps. Tissue was dispersed with 10 ml stop buffer (buffer A plus 12.5 µM CaCl₂, 10% FBS) with a transfer pipette, then passed through a 200 µm filter. Myocytes were settled by gravity for 10 min, supernatant removed, and pellet resuspended in 10 ml stop buffer. Ca²⁺was gradually increased from 12.5 µM to 0.25 mM through four additions of CaCl₂ over the course of 8 min. Myocytes were settled by gravity for 10 min, supernatant removed, and pellet resuspended in 5 ml normoxic buffer (buffer A plus 0.25 mM CaCl₂, 10% FBS). Myocytes were counted and viability determined by Trypan blue exclusion. One mouse heart typically yielded 1.2 × 10⁶ cells (75% viable) and cells were used for simulated IR injury immediately after isolation.

Mouse cardiomyocyte simulated IR injury

Myocytes were equilibrated at a concentration of 10⁵ cells/ml for 10 min in normoxic buffer. Incubations were in 50 ml round-bottom tubes in a shaking water bath (80 cycles/min) at 37 °C. Myocytes were divided in to the following groups (i) Control, 100 min in normoxic buffer; (ii) Simulated ischemia-reperfusion (SIR), 10 min normoxic buffer, then 1 h ischemia buffer (118 mM NaCl, 4.7 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 10 mM HEPES, 24 mM NaHCO₃, 0.25 mM CaCl₂, 10% FBS, gassed with 95% N₂, 5% CO₂, pH 6.5, 37 °C), followed by 30 min normoxic buffer; (iii) Drug + SIR, comprising treatment with compounds 10 min prior to SIR in normoxic buffer, then 1 h ischemia buffer, 30 min normoxic buffer. Drugs tested (final concentrations in parentheses) were: NS1619 (5 µM), NS11021 (500 nM) and diazoxide (10 µM). Drugs were dissolved in DMSO and the vehicle comprised <0.1% of total incubation volume. DMSO was included in the control and SIR groups at the appropriate times. To change buffers, cells were centrifuged at 31 × g for 2 min, and the pellet resuspended in the appropriate buffer. At the end of protocols, viability was determined by Trypan blue exclusion.

Isolated mitochondrial BK channel activity

Mitochondria were isolated from 3 mouse hearts as previously described  (Wojtovich et al., 2010). The mitochondria were incubated with 20 µM BTC-AM and 0.05% Pluronic F-127 for 10 min at room temperature. The final mitochondrial pellet was suspended in 225 µl and stored on ice until use, within 1.5 h Tl⁺ uptake into mitochondria was measured using a Varian Cary Eclipse spectrofluorimeter as previously described (Wojtovich et al., 2010) by monitoring changes in BTC fluorescence (λ_EX 488 nm, λ_EM 525 nm). Where indicated, ATP (to block mK_ATP channel), NS compounds (to open mBK channels) and paxilline (to block BK channels) were present at concentrations indicated.

Total heart histochemical acetylcholinesterase staining

Following IR protocols as described above, hearts were perfused with 10 mM phosphate buffered saline (PBS; pH 7.4 at 25 °C) for 3 min at 4 ml/min. Control hearts not exposed to IR injury were cannulated and perfused with PBS. Hearts were then stained for acetylcholinesterase (Rysevaite et al., 2011). Hearts were prefixed with 4% formalin for 30 min at 4 °C, then washed with PBS and incubated in PBS supplemented with hyaluronidase (0.5 mg/100 ml) for 18 h at 4 °C. The hearts were then placed in Karnovsky-Roots medium (60 mM sodium acetate, 2 mM acetylthiocholine iodide, 15 mM sodium citrate, 3 mM CuSO₄, 0.5 mM K₃Fe(CN)₆, 0.1% trition-X 100, 0.5 mg/100 ml hyaluronidase, pH 5.6 at 4 °C) for 3 h at 4 °C. Following staining, hearts were kept in 10% formalin and visualized under fluorescent illumination using GFP filters on an Olympus MVX wide field dissecting microscope. Images were obtained using a PixeLink camera. The fluorescent illumination provided increased contrast compared to visible light with the monochrome CCD.

Statistics

Data presented are means ± SEM. Statistical significance (p < 0.05) between multiple groups was determined using analysis of variance (ANOVA) with Bonferroni correction. Mixed effects models were generated using “R” software (R Development Core Team) and the R packages lme4 (Bates & Maechler, 2009) and languageR (Baayen, 2008; Baayen, 2009). As fixed effects, we included treatment conditions, genotype, and their interaction in the model. Normality and homogeneity were checked using visual inspections of plots of residuals against fitted values. Likelihood ratio tests comparing the models with fixed effects to the null models using only random effects, were used to assess the accuracy of the mixed effects models. The model was further analyzed using pair wise comparison t-test. We rejected results where the model including fixed effects did not differ significantly from the null model. For all mixed effects models we present MCMC-estimated p-values that are considered significant at the α < 0.05 level.

Protection of intact hearts by NS1619 and NS11021 requires Slo1

Previously we showed that cardioprotection by both IPC and APC in intact hearts is SLO1 independent (Wojtovich et al., 2011). However, the SLO1 openers NS1619 and NS11021 are still reported to protect against IR injury in intact hearts  (Bentzen et al., 2009; Xu et al., 2002), suggesting that SLO1 alone can afford protection. To examine the pharmaco-genomics of SLO1, NS1619 and NS11021 in more detail, we utilized an ex-vivo perfused heart model of IR injury. Representative raw data tracings of cardiac function during IR injury are shown in Fig. 1. Hearts from WT and Slo1^−/− mice exhibited similar sensitivity to IR (quantitative data shown in Fig. 2), and as previously reported both NS1619 and NS11021 protected the WT heart from IR injury (enhanced functional recovery and reduced infarct size), with NS11021 being 10-fold more potent than NS1619. As anticipated, protection was lost in Slo1^−/− hearts, suggesting that SLO1 is a bona-fide target for the cardioprotective effects of NS1619 and NS11021 in the intact heart. While it is possible that SLO1 is merely a downstream mediator of protection by NS1619 and NS11021, Occam’s razor would suggest SLO1 is the actual target.

Protection of isolated cardiomyocytes by NS1619 and NS11021 is Slo1 independent

Slo1 mRNA is expressed at low levels in whole heart (Chen et al., 2005; Jiang et al., 1999; Tseng-Crank et al., 1994), and it has been suggested that SLO1 is present in cardiomyocyte mitochondria (Xu et al., 2002). We therefore hypothesized that SLO1 in cardiomyocyte mitochondria may play a role in mediating the protective effects of NS1619 and NS11021. To test this hypothesis, adult mouse ventricular cardiomyocytes were isolated from WT and Slo1^−/− hearts and subjected to simulated IR injury. As expected, both NS1619 and NS11021 elicited cytoprotection (reduced cell death) in response to IR (Fig. 3). However, this protection was present in both WT and Slo1^−/− derived myocytes, indicating it did not require SLO1 activity. As a control, the mitochondrial K_ATP opener diazoxide was protective in both WT and Slo1^−/− cells. Overall these data suggest that the protection of cardiomyocytes by NS1619 and NS11021 may be due to off-target (non-SLO1) effects of these molecules, as reported elsewhere (Park et al., 2007; Saleh et al., 2007; Holland et al., 1996; Aldakkak et al., 2010; Aon et al., 2010; Cancherini, Queliconi & Kowaltowski, 2007; Debska et al., 2003; Kicinska & Szewczyk, 2004).

NS1619 and NS11021 exhibit Slo1 independent effects on isolated mitochondria

To directly test whether NS1619 and NS11021 were able to activate BK channels in isolated mitochondria, a modified thallium (Tl⁺) flux assay was utilized (Wojtovich et al., 2010). As expected, both compounds were able to elicit a Tl⁺ flux, in a manner sensitive to the BK channel blocker paxilline (Fig. 4). However, this flux was evident in mitochondria derived from both WT and Slo1^−/− hearts, indicating it did not originate at the level of SLO1. Furthermore, bona fide mitochondrial K⁺ channels exhibit a slowing of Tl⁺ flux kinetics by increasing concentrations of K⁺ and not Na⁺ (since only K⁺ competes for Tl⁺ uptake). However, such competition was not seen for NS induced Tl⁺ flux (data not shown), suggesting it is not ion-specific. Together these data support the notion that NS compounds have non-specific effects on mitochondria. Such effects may account for the non Slo1-dependent protection observed in isolated cardiomyocytes (Fig. 3).

SLO1 dependent protection in intact hearts requires cardiac neuronal function

So far, our data show that SLO1 is not required for protection by NS1619 and NS11021 in isolated cardiomyocytes (Fig. 3), but is required for protection by these compounds in the intact heart (Figs. 1 and 2). This suggests that in the intact heart, protection occurs via SLO1 in a mechanism that is non-cardiomyocyte autonomous. Several non-myocyte cardioprotective mechanisms exist, including remote preconditioning which is thought to comprise a neuronal signaling component (Hausenloy & Yellon, 2008). We therefore sought to investigate a role for SLO1 in cardioprotection via non myocyte signaling, with a focus on intrinsic cardiac neurons. Importantly, the ex-vivo heart preparation used herein is denervated, such that surviving neurons are not central but rather they are intrinsic to the heart. In addition SLO1 channel activity has been described in cardiac neurons (Franciolini et al., 2001) and Purkinje fibers  (Callewaert, Vereecke & Carmeliet, 1986), where it plays a role in regulating heart rate (Wojtovich et al., 2011; Imlach et al., 2010).

Figure 5 shows that blockade of neuronal function with the synaptic transmission inhibitor hexamethonium, prevented cardioprotection (reduction in infarct size) by NS1619 in intact hearts. This effect was specific to SLO1 and NS1619, since hexamethonium had no effect on protection afforded by the mitochondrial K_ATP channel activator atpenin A5 (Wojtovich & Brookes, 2009). These data also indicate that NS1619 protects via a mechanism that is distinct from mitochondrial K_ATP channel opening.

To further investigate the role of cardiac neurons in protection afforded by SLO1 activators, we assessed neuronal status by histochemical staining for acetylcholinesterase activity (AChE). No differences in AChE staining were found between control, IR, and IR + NS1619 hearts, suggesting the acute time frame of IR injury in this model was not sufficient to precipitate neuronal loss (Fig. 6A–C). We therefore hypothesized that a more direct test of cardiac neuronal function might be necessary to assess neuronal protection in this acute model. Hence, the response of heart rate to nicotine (which induces transient tachycardia via nicotinic Ach receptors (Tosaka et al., 2003; Haass & Kubler, 1997)) was measured after IR injury. As seen in Fig. 6D, no response to nicotine was seen in control hearts subjected to IR, indicating loss of neuronal function. However, hearts protected with NS1619 exhibited a robust post-IR nicotine response, and this effect was blocked by pre-treatment with hexamethonium, indicating a preservation of nicotine-sensitive intrinsic cardiac neurons. Notably, hearts protected with the mK_ATP opener atpenin A5 did not exhibit a preserved nicotine response. This is consistent with the data in Fig. 4 showing that even though atpenin A5 was protective, this cardioprotection was insensitive to hexamethonium. Overall these data indicate that both mK_ATP and SLO1 channels can protect the heart (Figs. 1, 2 and 5), but they do so via distinct mechanisms, the latter requiring cardiac neuron function.